An Integration of Genetic Feature Selector, Histogram-Based Outlier Score, and Deep Learning for Wind Turbine Power Prediction

ABSTRACT

During the last decades, the importance of clean energy resources is being increased. Wind is one of the most significant clean energy resources. Forecasting the output power of wind turbines is important for turbine control to improve power grids’ performance and maintenance. In this study, a novel method for predicting the power of wind turbines is proposed based on integrating data preprocessing, re-sampling, feature selection (genetic algorithm), and outlier detection (Histogram-Based Outlier Score) techniques to prepare the data for the deep learning (DL) algorithms. The results show that, after removing features chosen by the genetic algorithm (GA) method, the mean absolute error (MAE) reduced considerably to 333.7. Integrating Histogram-Based Outlier Score (HBOS) with genetic algorithm (GA) significantly decreased the MAE to 488. Comparing the results with benchmark machine learning algorithms, namely Light Gradient Boosting Machine (LGBM), Extreme Gradient Boosting Regression (×GBR), K-Nearest Neighbor (KNN), Support Vector Regression (SVR), and Recurrent Neural Networks (RNN) models, shows a remarkable improvement in the accuracy of turbine power prediction for about 78.7, 944.9, 104.7, 1456.6, and 17.1 in mean absolute error (MAE), respectively.

KEYWORDS:

• Wind turbine
• power prediction
• machine learning
• deep learning
• genetic algorithm

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. http://www.sfpgmc.co.ir/en/projects/bina.

Additional information

Notes on contributors

Parastou Fahim

Parastou Fahim obtained her B.Sc in Electrical Engineering from Shahid Beheshti University. She also completed her M.Sc in Electrical Engineering at the Ferdowsi University of Mashhad.

Nima Vaezi

Nima Vaezi obtained his B.Sc in Electrical Engineering from Sadjad Technology University. He also completed his M.Sc in Electrical Engineering at the Ferdowsi University of Mashhad. He is interested in applying Artificial Intelligence/Deep Learning in renewable energy, medicine, and power systems.

Amin Shahraki

Dr Amin Shahraki received his Ph.D from the University of Oslo, Norway, in 2020. He was a postdoctoral researcher at the Communication Systems Division, Fraunhofer IIS, Erlangen, Germany. Now, he works at ABB Corporate Research Germany as a research scientist focusing on Industrial networking. He is also the scientific reviewer of more than 50 high-reputation journals. His current research interests are Internet of Things, Machine Learning for network management, cognitive communications and networking, network behavior analysis, and time series analysis.

Mohammad Khoshnevisan

Professor Mohammad Khoshnevisan is an Affiliated Research Professor at Northeastern University, Physics Department, Boston, USA. He earned his BA in Mathematics from Christian Brothers University in the United States. He obtained his PhD from the University of Melbourne, Department of Computer Science and Software Engineering, Australia. He worked as a tenured lecturer, senior lecturer, and associate professor in Australia and the GCC region for approximately 17 years. He also gives lectures to engineering students for a graduate-level course in Advanced Engineering Mathematics and an undergraduate course in Artificial Intelligence. Professor Khoshnevisan has developed various Artificial Intelligence applications in Physics, Medicine, and Engineering with his research team and Professor Kosrow Dehnad at Columbia University in the United States. His research focus is on applying Quantum Mechanics in Science and Engineering and the application of Mathematical Modeling and Deep Learning/Machine Learning in Physics and Engineering. He was formally invited as a visiting scholar by UC Berkeley and Harvard University. He is a PhD examiner for Deakin University in Australia. He was a reviewer for Advances in Space Research (Q1 in Aerospace Engineering) and the International Federation of Automatic Control (IFAC) World Congress. He has received a Certificate of Achievement for his contribution to the BISC-FLINT-CIBI international joint workshop on soft computing for internet and bioinformatics from the world-renowned scientist and inventor of Fuzzy Logic, late Professor Emeritus Lotfi A. Zadeh, University of California-Berkeley. Professor Khoshnevisan has published joint papers in the International Journal of Advanced Manufacturing Technology and the International Journal of Intelligent Transportation Systems Research.